Interface remains down

[BikeMan]

Dear all,

I get an issue this week with a cluster xl cluster. We are running R80.20T183 on 5400 appliance. Here is the story:

M1 is active.

Disconnect M1-eth1 (it contains several vlan interfaces) from M1, M2 becomes active;

Reconnect M1-eth1: eth1 remains in error state with M1 status DOWN;

In this case, disconnecting M2-eth1 will break all connections.

The only way to have M1-eth1 up is to reboot the aplliance.

This is not what is expected.

Working with ClusterXL_admin script is working fine.

Any idea (other than upgrade...) ?

Thanks for your help.

[Chris_Atkinson]

Have you verified that you can see ARP via this interface prior to rebooting?

What are the portfast settings for this interface on the switch side?

Is this an onboard NIC or expansion card?

Note R80.20 is End of Support.

[BikeMan]

Hi,

Thanks for answering.

No, didn't check ARP table but I guess it was empty because interface was down.

Didn't check switch setting yet because I want to be sure that it is not a CP issue first. But will check shortly.

These are onboard NIC.

Yes, R80.20 is out of support but that does not mean the interface has to stay down.

Rgds,

[Chris_Atkinson]

Ultimately trying to clarify was the interface link-state still down upon cable reconnection or only marked down by CXL due to no CCP flow at the time?

Hopefully it doesn't become a problem here but at some point the status of R80.20 may impede timely resolution to support tickets.

[Lesley]

Hello,

Do you have the output of the following commands during the issue? Is the problematic interface by any chance the highest or lowest VLAN number of this interface? Maybe filter out the 'private' info before you share the output 🙂 

cphaprob stat

cphaprob -a if

cphaprob -l list

cpstat ha -f all | more

[BikeMan]

Hello,

No I don't have. It happens nigthly while I was sleeping, the local IT had to reboot.

Rgds,

[the_rock]

Just an idea...I wont even comment on version being unsupported, as it makes no difference, same command would apply to even R55 lol

So, can you run from expert mode the following -> grep -i DOWN /var/log/messages*

This would give you all lines with that word in all the messages files. Not sure if it may provide better info as to why its happening, but worth a try.